Proc. Natl. Acad. Sci.

USA
Vol. 96, pp. 7604–7609, June 1999
Anthropology

The early Upper Paleolithic human skeleton from the Abrigo do

Lagar Velho (Portugal) and modern human emergence in Iberia
(Neandertalsymandibleypostcraniaydentitionyradiocarbon dating)

CIDÁLIA DUARTE*†, JOÃO MAURÍCIO‡, PAUL B. PETTITT§, PEDRO SOUTO‡, ERIK TRINKAUS¶i**,
HANS VAN DER PLICHT††, AND JOÃO ZILHÃO‡‡
*Instituto Português do Património Arquitectónico, Divisão de Conservação e Restauro, Palácio da Ajuda, 1400-206 Lisbon, Portugal; †Department of
Anthropology, 13-15 Henry Marshall Tory Building, University of Alberta, Edmonton T6G 2H4, AB, Canada; ‡Sociedade Torrejana de Espeleologia e
Arqueologia, Quinta da Lezı́ria, 2350 Torres Novas, Portugal; §Radiocarbon Accelerator Unit, Research Laboratory for Archaeology and the History of Art,
University of Oxford, 6 Keble Road, Oxford OX1 3QJ, England; ¶Department of Anthropology, Campus Box 1114, Washington University, St. Louis, MO 63130;
i
Unité Mixte de Recherche 5809 du Centre National de la Recherche Scientifique, Laboratoire d’Anthropologie, Université de Bordeaux I, 33405 Talence, France;
††Centrum voor Isotopen Onderzoek, Faculteit der Wiskunde en Natuurwetenschappen, Rijksuniversiteit Groningen, Nijenborgh 4, 9747 AG Groningen, The
Netherlands; and ‡‡Instituto Português de Arqueologia, Avenida da India 136, 1300 Lisbon, Portugal

Contributed by Erik Trinkaus, April 26, 1999

ABSTRACT The discovery of an early Upper Paleolithic
human burial at the Abrigo do Lagar Velho, Portugal, has
provided evidence of early modern humans from southern
Iberia. The remains, the largely complete skeleton of a '4-
year-old child buried with pierced shell and red ochre, is dated
to ca. 24,500 years B.P. The cranium, mandible, dentition, and
postcrania present a mosaic of European early modern human
and Neandertal features. The temporal bone has an interme-
diate-sized juxtamastoid eminence. The mandibular mentum
osseum and the dental size and proportions, supported by
mandibular ramal features, radial tuberosity orientation, and
diaphyseal curvature, as well as the pubic proportions align
the skeleton with early modern humans. Body proportions,
ref lected in femorotibial lengths and diaphyseal robusticity
plus tibial condylar displacement, as well as mandibular
symphyseal retreat and thoracohumeral muscle insertions,
align the skeleton with the Neandertals. This morphological
mosaic indicates admixture between regional Neandertals and
early modern humans dispersing into southern Iberia. It
establishes the complexities of the Late Pleistocene emergence
of modern humans and refutes strict replacement models of
modern human origins.
During the past decade it has become apparent that the human
biological and cultural evolutionary transitions between late
archaic (Neandertal) and early modern humans and between
the Middle and Upper Paleolithic occurred relatively late
throughout most of Iberia (1–4). It is now certain that the
Middle Paleolithic of most of Iberia south of the Pyrenees
lasted until about 30,000 years B.P., perhaps as late as 28,000
years B.P., and that the initial stages of the Upper Paleolithic
known to have occurred further north (the Châtelperronian
and the early Aurignacian) were never present in this region.
Moreover, human paleontological evidence from Zafarraya in
southeastern Spain indicates that late Middle Paleolithic tech-
nology from this region was the product of Neandertal pop-
ulations (5). This evidence has led to hypotheses as to why
Middle Paleolithic Neandertals endured in the cul-de-sac of
Iberia between 5,000 and 10,000 years after they had been
replaced (by whatever historical processes) elsewhere in Eu-
rope (4, 6, 7). The ‘‘Ebro Frontier’’ model (3, 6–8) suggests
that the basin of the Ebro river of northern Spain represented
a biogeographical and ecological barrier to the diffusion of the
Upper Paleolithic innovations developed by late Neandertal
populations to the north (the Châtelperronian) and, subse-
quently, to the dispersal of the first modern human groups in
western Europe.
PNAS is available online at www.pnas.org.
Even though both the late Middle Paleolithic and early
Upper Paleolithic are increasingly well known and chronolog-
ically situated south of the ‘‘Ebro Frontier’’ (4, 6, 9), diagnostic
human remains associated with early Upper Paleolithic indus-
tries in this region have been elusive. The discovery of a largely
intact early Upper Paleolithic child’s burial at the Abrigo do
Lagar Velho in Portugal therefore adds significantly to our
knowledge of the biology and burial practices of the earliest
Iberian modern humans and sheds light on the nature of the
transition from Neandertals to their successors in one of the
last archaic human frontiers.
THE ABRIGO DO LAGAR VELHO
The archeological site consists of deposits within horizontal
fissures and along the current base of a limestone cliff on the
south side of the Lapedo Valley near Leiria, central western
Portugal (lat 39° 459 250 N, long 8° 439 580 W). The site was
damaged by earth removal in 1992, exposing an Upper Paleo-
lithic sequence and coming within a few centimeters of the
burial along the base of the cliff. On November 28, 1998, the
site was discovered by J.M. and P.S., who also found the left
hand and forearm bones of the child in a burrow. The following
week C.D. and J.Z. confirmed the presence of Paleolithic
deposits and a human burial. A salvage excavation directed by
J.Z. ensued from December 12 to January 7, 1999, in the
framework of which the burial was excavated by C.D. Paleon-
tological analysis by C.D. and E.T. commenced on January 4,
1999.
The child’s skeleton, Lagar Velho 1 (Fig. 1), was on its back
parallel to the cliff base, with the head to the east and left side
against the cliff. The cranium and mandible were damaged by
earth removal, but the preservation of the left temporal bone
and mandible indicates that the head had tilted toward the cliff
face. The collapse and folding in situ of the right ribs suggest
that the thorax was similarly tilted. The pelvis was horizontal,
and the feet were plantar-flexed and crossed. The skeleton and
the containing sediment are heavily stained with red ochre, but
the alteration of the sediment stopped at the outer border of
the skeleton, suggesting a wrap around the body. Vertically
oriented animal bones and stones outlined the ochre-stained
sediments; radiocarbon dating of one such bone yielded a
result that is consistent with the hypothesis that these represent
intentional burial features and not a natural disposition of
deposit components caused by digging a burial pit.
A Commentary on this article begins on page 7117.
**To whom reprint requests should be addressed. e-mail: trinkaus@
artsci.wustl.edu.

7604
 Anthropology: Duarte et al. Proc. Natl. Acad. Sci. USA 96 (1999) 7605

vertebral column [23,920 6 220 years B.P. (OxA-8422)] im-

mediately overlying the legs. The burial therefore occurred
between 24,000 and 25,000 years B.P. and probably between
24,500 and 25,000 years B.P.

AGE AND BODY SIZE

The remains are those of a juvenile (Figs. 1 & 2). All of the
deciduous teeth (right di2 to left dm2 plus left dc1 and dm2)
have apically closed roots, indicating a median minimum age
of '3.0 years. The I2 is almost at Crc; the C1 is 'C3y4; the P4
is 'Cco; the M1 and M1 are 'R1y4; and the M2 is 'Coc
(nomenclature established by Moorrees et al.; ref. 14). The
levels of calcification for the latter four teeth provide median
ages (maleyfemale) of 3.4y3.5, 3.9y3.9, 4.9y4.8, and 4.9y4.5
years, by using Euro-American standards (15). The average
estimate is 4.2 years postnatal, and the probable range is
3.5–5.0 years.
The more complete right femur has an intermetaphyseal
length of 198 mm plus proximal and distal epiphyseal thick-
nesses of 10.0 mm and 13.3 mm, respectively. Adding these and
rounding up because of the absence of the epiphyseal cartilages
provides an interepiphyseal length of '225 mm. Estimates
from the ratios of 4- to 18-year-old modern Euro-American
mean femoral lengths (16) predict adult femoral lengths of
'450 mm (male) and '410 mm (female). These values are
below the means for adult European earlier Upper Paleolithic
males (478.0 6 21.4 mm; n 5 13) and females (427.3 6 19.0
mm; n 5 9) but within their ranges of variation; the values are
close to the means for Neandertal males (443.3 6 18.7 mm; n 5
11) and females (400.3 6 14.2 mm; n 5 3).

MORPHOLOGICAL CONSIDERATIONS
FIG. 1. Lagar Velho 1 in situ, with damaged skull and left forearm
elements already removed. Materials and Methods. This preliminary assessment of
Lagar Velho 1 is concerned with its morphological affinities to
The only diagnostic archeological item in the burial was a northwestern Old World late archaic humans (Neandertals)
pierced Littorina obtusata shell found near the cervical verte- versus early modern humans. The latter sample consists pre-
brae; it is identical to those from Level Jb of the nearby site of dominantly of Aurignacian and Gravettian remains between
Gruta do Caldeirão (Tomar) dated to 26,020 6 320 years B.P. 20,000 and 30,000 years B.P. Given the dearth of earlier Upper
(OxA-5542; ref. 6). Similar burials with pierced shells andyor Paleolithic juvenile human remains, the comparisons also
teeth and a covering of ochre are known particularly from the involve Near Eastern Middle Paleolithic (Qafzeh-Skhul) early
Gravettian of Europe, especially from Britain (Paviland), Italy modern humans. Recent human comparative data are in-
(Arene Candide, Barma Grande, Caviglione, Ostuni), and the cluded as appropriate, the samples deriving from temperate
Czech Republic (Brno, Dolnı́ Věstonice; refs. 10–12). European and North American samples (17–19). The relative
The uppermost 2–3 m of the shelter’s fill were largely positions of the fossil specimens in the postcranial propor-
removed and current ground level represents the surface of an tional assessments are based on distributions of raw residuals
estimated 3- to 4-m sequence of early Upper and possibly from the reduced major axis lines for the recent human
Middle Paleolithic strata. A '60-cm thick hanging remnant is samples and are expressed as z scores.
preserved within a fissure running along the back wall of the
shelter and contains a Proto-Solutrean (level 6) through
Middle Solutrean (level 9) sequence. In Portugal, assemblages
with diagnostic laurel-leaf point fragments date to ca. 20,000–
20,500 years B.P., and the Proto-Solutrean dates to ca. 21,500
years B.P. (6, 13). Radiocarbon dating of charcoal from level
9 yielded a date of 20,200 6 180 years B.P. (OxA-8419), and
samples of charcoal from level 6 yielded results of 21,180 6 240
years B.P. (OxA-8420) and 21,380 6 810 years B.P. (Sac-1561);
a stratigraphically less reliable level 6 sample yielded 22,180 6
180 years B.P. (OxA-8418).
The position of the burial, '2.5 m below the Proto-
Solutrean remnant, as well as its archeological resemblances to
Gravettian burials elsewhere, suggests an age several millennia
earlier. This inference is supported by accelerator mass spec-
trometry radiocarbon dating of charcoal [24,860 6 200 years
B.P. (GrA-13310)] and Cervus elaphus remains [24,660 6 260 FIG. 2. Lingual view of the mandible and dentition, showing the
years B.P. (OxA-8421) and 24,520 6 240 years B.P. (OxA- degrees of dental development, the symphyseal retreat, and the
8423)] directly associated with the burial and of a vertebra prominent right tuberculum laterale. (Black bar and white bar 5 1 cm
from a semiarticulated section of a Oryctolagus cuniculus each.)
7606 Anthropology: Duarte et al.
FIG. 3. Anterior view of the mandibular symphysis showing the
development of the mentum osseum. Symphyseal height 5 24.7 mm.
The Temporal Bone. The mastoid region in the left temporal
bone is preserved with little damage. It has a clear juxtama-
stoid eminence, which extends to the same level as the mastoid
process tip. The relative sizes of these two processes is inter-
mediate between Neandertal juveniles with their larger jux-
tamastoid eminences (20) and early modern humans with their
more projecting mastoid processes (21, 22).
The Mandible. The anterior mandibular symphysis has the
full development of a mentum osseum (Fig. 3). There is a very
prominent tuber symphyseos with an exceptionally protruding
tuberculum laterale on each side. These combine to create a
deeply excavated incisura mandibulae anterior. The midline of
the tuber symphyseos extends superiorly toward the alveoli. By
any criteria, this symphyseal configuration is a prominent
development of a ‘‘chin,’’ clearly distinct from the modest
development of these elements in similarly aged Neandertals
(19, 23) and pronounced even for a developmentally compa-
rable modern human (21, 24).
This anterior symphyseal configuration is combined with a
posterior retreat of the anterior symphysis, indicated by an
anterior symphyseal angle (of the alveolar plane to the infra-
dentale–pogonion line) of 89°. It is close to the mean for early
juvenile Neandertals (90.3° 6 4.4°; n 5 7) and below that of
similarly aged recent humans (101.5° 6 4.7°; n 5 15; ref. 18).
The corpus and ramus have a modest planum alveolare, a
mental foramen below the dm1, an open (V shaped) mandib-
ular foramen, a symmetrical mandibular incisure, and an
absence of a superior medial pterygoid tubercle (despite
rugosity for the medial pterygoid muscle). These features
either align the specimen with recent humans or do not
discriminate between the two groups (25).
The Dentition. The deciduous dental remains have no
macroscopic pathological alterations and minimal incisor at-
trition. The first molar crown diameters (Table 1) are similar
to those of Neandertals and early modern humans, who are not
significantly different in these dimensions. The I2 breadth
shows a gradual decrease from Neandertals to Qafzeh-Skhul
humans to earlier Upper Paleolithic humans to recent Euro-
peans, and Lagar Velho 1 falls between the means of the last
two samples. Late archaic and early modern humans differ
Table 1. Comparative permanent dental metrics
Sample I2 breadth, mm M1 length, mm
Lagar Velho 1 (6.6)* 10.5
Neandertals 7.8 6 0.4 [27] 11.2 6 0.7 [26]
Qafzeh-Skhul 7.2 6 0.6 [8] 11.3 6 0.6 [13]
Early Upper Paleolithic 6.9 6 0.5 [22] 10.8 6 0.8 [29]
Recent Europeans 6.3 6 0.4[106] 10.0 6 0.6[104]
Data are shown as means 6 SD [n].
*Includes estimated additional crown formation, as is indicated by the parentheses.
Proc. Natl. Acad. Sci. USA 96 (1999)
more in their anterior-to-posterior dental proportions than in
absolute dental dimensions (26), and the relative I2 versus M1
breadths of Lagar Velho 1 fall with the early modern humans
and separate from the Neandertals.
The Postcranial Remains. At least one of each of the major
long bones was preserved in the appendicular skeleton of
Lagar Velho 1, with only the radius lacking a complete
intermetaphyseal length (Fig. 4). The intermetaphyseal max-
imum lengths (left side for all but the femur) are, for the
humerus, 143 mm; for the ulna, 122.5 mm; for the femur, 198
mm; and for the tibia, 155 mm. The clavicles are not sufficiently
preserved to provide lengths, and major portions of the pelvis
remain.
The right femur has a neck-shaft angle of '128°. This value
is similar to the values of juvenile (Roc de Marsal 1, 130°; La
Ferrassie 6, 132°) and early adolescent (Teshik-Tash 1, 128°)
Neandertals and the juvenile Skhul 1 early modern human
(131°). Neck-shaft angles decrease with normal locomotor
loading during development (27), such that (based on a recent
Native American sample; ref. 28) it should decrease '2%
between the juvenile and adult years. The predicted adult value
for Lagar Velho 1 (125°) is similar to the values of Neandertals
(121.0° 6 4.7; n 5 9) and earlier Upper Paleolithic Europeans
(118.8° 6 6.0; n 5 13) but below the Qafzeh-Skhul values
(133.2° 6 2.6; n 5 6). The right femur has a metaphyseal
bicondylar angle (29) of 8°. The greater medial distal epiph-
yseal thickness (13.3 mm versus 12.7 mm) suggests that the
articular bicondylar angle was slightly higher. These values are
attained by recent humans at about 4–5 years of age (29). This
pattern is echoed in the proximal tibial metaphyseal orienta-
tion; the medial metaphyseal retroversion angle (12°) is close
to articular angles of mature Neandertal (15.0° 6 2.2; n 5 5)
and earlier Upper Paleolithic (15.8° 6 5.3; n 5 8) tibiae.
These angles indicate that locomotor loading levels were
similar to those of European Late Pleistocene hominids and
nonurban recent humans (29, 30). In combination with the
apparent absence of pathological lesions on the dental and leg
remains, these angles imply normal and active locomotor
behavior for Lagar Velho 1 in the context of little or no
systemic developmental stress. As a result, it is possible to
consider other proportions of its lower limbs as reflections of
normal skeletal morphology.
Early and recent modern humans are largely distinct from
archaic humans in the relative mediolateral shortness of their
pubic bones (31), and this contrast appears early in develop-
ment (18). A comparison of pubic (acetabulosymphyseal)
breadth to femoral length shows a regular pattern of devel-
opment (r2 5 0.857) with the La Ferrassie 6 Neandertal falling
significantly above the recent human distribution (Table 2).
The Lagar Velho 1 pubic length is within the recent human
distribution but is still relatively high for a modern human.
Neandertals and European early modern humans contrast in
the length of the tibia relative to the femur (32). Moreover,
differences in this ecogeographically patterned body propor-
tion (in which the Neandertals are hyperarctic and the earlier
Upper Paleolithic humans subtropical) appear early in devel-
opment among recent humans (33, 34) and Neandertals (18,
35–37) and remain stable over multiple millennia (32). Tibial
metaphyseal length versus femoral diaphyseal intermetaphy-
M1 breadth, mm M1 length, mm M1 breadth, mm
11.8 12.0 10.9
12.0 6 0.8 [26] 11.4 6 0.7 [44] 10.9 6 0.5 [44]
12.2 6 0.7 [13] 11.7 6 0.8 [8] 11.4 6 0.6 [9]
12.1 6 0.8 [29] 11.5 6 0.8 [35] 11.0 6 0.9 [38]
11.3 6 0.5[104] 10.7 6 0.6[106] 10.3 6 0.5[106]
 Anthropology: Duarte et al.
FIG. 4. Anterior view of the long bones. (Black and white bars 5
1 cm each.)
seal length provides a tight linear fit between the variables for
the age range ('2–8 years) in a cool-temperate modern human
sample (r2 5 0.983), with Skhul 1 falling above the line and
both La Ferrassie 6 and Lagar Velho 1 falling significantly
below the recent humans (Table 2). Given the hyperarctic
versus tropical proportions of Neandertal versus Qafzeh-Skhul
adult remains (38), the separation of the La Ferrassie and
Skhul specimens is expected. However, the low position of
Lagar Velho 1 is unexpected, given the relatively long tibiae of
all known European earlier Upper Paleolithic humans (32).
These contrasts in adult limb-segment proportions are com-
bined with the Neandertals having very broad trunks and the
Qafzeh-Skhul and European earlier Upper Paleolithic trunks
being quite slender (32, 38). As a result of the contrasting body
masses (and hence baseline weight-bearing loads) relative to
femoral and especially tibial lengths, Neandertals seem to have
robust leg diaphyses, whereas these early modern humans
seem gracile (39). The apparent differences in structural
hypertrophy, however, disappear once body proportions are
taken into account (40–43). Consequently, given that mature
Table 2. z scores based on raw residuals from the reduced major axis line of the recent human cool-temperate sample for immature Lagar
Velho 1, Neandertal, and Qafzeh-Skhul early modern human femora and tibiae
Sample Public breadthyfemur length Tibiayfemur length
Lagar Velho 1 1.247 22.044
Neandertals
La Ferrassie 6 2.119 22.415
Roc de Marsal 1 — —
Teshik-Tash 1 — —
Early modern humans
Qafzeh 10 — —
Skhul 1 — 0.493
A z score . 1.95 indicates that P , 0.05.
Proc. Natl. Acad. Sci. USA 96 (1999) 7607
diaphyseal hypertrophy reflects in large part developmental
trajectories (44), it is expected that immature Neandertals and
early modern humans had similar femoral and tibial diaphyseal
hypertrophy once body proportions were incorporated into the
analysis. In other words, any significant differences in diaph-
yseal hypertrophy should primarily reflect contrasts in body
proportions (relative trunk size for the femur and relative
trunk size plus leg segment proportions for the tibia) and only
secondarily differences in lower limb robusticity.
In the comparison of femoral midshaft circumference to
length (Fig. 5), Lagar Velho 1 and the three Neandertal
specimens (La Ferrassie 6, Roc de Marsal 1, and Teshik-Tash
1) fall significantly above the recent human distribution (Table
2). The Qafzeh-Skhul early modern humans are more modest
in their positions. In the tibial comparison (Fig. 5), the early
modern Skhul 1 is in the middle of the recent human distri-
bution (reflecting the competing effects of linear body pro-
portions and Pleistocene Homo hypertrophy), and La Fer-
rassie 6 is well above the recent human distribution (reflecting
the combination of hypertrophy and hyperarctic body propor-
tions). Lagar Velho 1 is beyond the Neandertal specimen,
clearly separate from the recent humans. Therefore, either
Lagar Velho 1 had exceptionally robust femora and tibiae, or
more likely, it shared the hyperarctic trunk and lower limb
proportions of the Neandertals.
The tibial plateau is dorsally displaced 20 mm from the
midcondyles to the anterior tibial tuberosity, a pattern seen
among the Neandertals and associated with increasing the
quadriceps femoris moment arm to compensate for elevated
body-mass to leg-length proportions (45). This pattern further
supports the inference of elevated relative trunk mass in Lagar
Velho 1.
The left humerus has clear diaphyseal torsion and a prom-
inent ridge along the pectoralis major insertion leading up to
the anterior greater tubercle. There is rugosity for the pecto-
ralis major attachment, and the ridge creates a marked inter-
tubercular sulcus and an anterolateral to posteromedial elon-
gation of the diaphyseal cross section. This proximal humeral
diaphyseal morphology implies hypertrophy of the thoracohu-
meral musculature, a pattern seen in Neandertals but usually
little developed among early modern humans (39, 46).
The radius has an interosseus crest in line with the dorsal
third of radial tuberosity. The resultant anteromedial position
of the radial tuberosity contrasts with the directly medial
position of most Neandertals, such that the orientation found
here is present in 26.9% of Neandertals (n 5 13) but in 100%
of European earlier Upper Paleolithic humans (n 5 23; refs.
47, 48). At the same time, the radius has little lateral curvature,
characteristic of early modern human radii but unlike most
mature and immature Neandertal radii (46, 49).
DISCUSSION
These morphological considerations provide a mixed impres-
sion of the Lagar Velho 1 juvenile early Upper Paleolithic
z Scores
Femur circumferenceylength Tibia circumferenceylength
2.815 2.582
2.501 2.402
2.032 —
2.643 —
1.646 —
1.148 0.037
7608 Anthropology: Duarte et al.
FIG. 5. Midshaft circumference versus femoral intermetaphyseal
diaphyseal length (Upper) and versus tibial intermetaphyseal length
(Lower). Black hexagon, Lagar Velho 1; gray pentagons, European
Neandertals; gray squares, Qafzeh-Skhul humans; open triangles,
cool-temperate recent humans.
human skeleton relative to northwestern Old World Neander-
tals and early modern humans. It is buried in a distinctively
early Upper Paleolithic pattern, but its morphology is a mosaic
of Neandertal and early modern human features.
The temporal juxtamastoid eminence development is inter-
mediate between these two groups. The mandibular mentum
osseum and dental proportions align it completely with the
early modern humans. This alignment is supported by man-
dibular ramal discrete traits, orientation of the radial tuber-
osity, radial curvature, and relative pubic breadth. The fem-
orotibial length proportions and lower limb hypertrophy (fem-
oral and tibial diaphyseal robusticity and tibial condylar
posterior displacement) indicate hyperarctic Neandertal body
proportions distinct from those of European earlier Upper
Paleolithic humans. In addition, the mandibular symphysis
retreats at an archaic angle, and the thoracohumeral muscle
insertions are most similar to those of the Neandertals.
The body proportions of Iberian Neandertals are not known,
but all European Neandertals, including the Châtelperronian
St. Césaire 1, have hyperarctic body proportions (41, 50). It is
therefore likely that the Iberian Neandertals either had similar
body proportions or, given the more temperate glacial climate
of coastal southern Iberia (6), had less extreme cold-adapted
proportions. All of the European earlier Upper Paleolithic
human remains have subtropical body proportions, whether
Proc. Natl. Acad. Sci. USA 96 (1999)
measured directly or inferred from femoral and tibial diaph-
yseal scaling (32, 42, 43). The more moderate oxygen-isotope
stage 3 climate of southern Iberia versus the climate known for
southwestern France or central Europe makes it unlikely that
the inferred cold-adapted body proportions of Lagar Velho 1
were the product of regional climatic selection. They can only
indicate affinities to the Neandertals.
Although some of these features are developmentally plas-
tic, the mandibular and dental patterns as well as the ecogeo-
graphically related body proportions seem to be resistant to
environmental perturbations and to be evolutionarily stable
over at least the millennia of concern. They are therefore
appropriate markers for assessing the ancestry of Lagar
Velho 1.
There are two logical explanations for the morphological
mosaic seen in Lagar Velho 1. It could represent the ancestral
(plesiomorphous) pattern for Late Pleistocene European hu-
mans, or it could be the result of admixture between European
late archaic and early modern humans. The first explanation
can be rejected by the presence of a uniquely derived (auta-
pomorphous) feature of modern humans, its mentum osseum
development, supported by its chronological position after the
presence of both Neandertals and early modern humans in
Europe.
The morphological mosaic of the Lagar Velho 1 child
therefore indicates admixture between early modern humans
spreading through Iberia and local Neandertal populations.
Such morphological mosaics, with character states distinctive
to each parental group (directional dominance), plus inter-
mediate (additive) or divergent (overdominance or under-
dominance) configurations, characterize hybrids between sub-
species, species, and genera of primates and carnivores (51–
55). It is not known whether any one mosaic pattern would be
expected, given specified ancestral morphological patterns,
nor is it known how stable the pattern is likely to have been
over time. However, the mosaic seen in Lagar Velho 1
conforms to the known patterns of hybridization.
Moreover, the dating of the burial to several millennia after
the probable period of transition from Neandertal to early
modern human in southern Iberia (28,000–30,000 years B.P.),
and hence the persistence of a morphological mosaic for that
time period, indicates that this child was not the result of a rare
Neandertalyearly modern interbreeding but the descendant of
extensively admixed populations. There does not seem to be a
better means of explaining why, in spite of sharing a similar
material culture and identical burial practices with the Gravet-
tian elsewhere in Europe, Lagar Velho 1 is anatomically so
distinct.
The presence of such admixture supports the hypothesis (56,
57) of variable admixture between early modern humans
dispersing into Europe and local Neandertal populations. It
runs counter to the interpretation (58, 59) that, in western
Europe, this transition involved the extinction without descent
of the Neandertals and their complete replacement by dis-
persing early modern humans with more elaborate technolog-
ical and sociocultural systems.
The broader implications of this finding are multiple. First,
it means that the degree of abruptness in sociocultural and
technological complexes around the time of the transition from
the Middle to Upper Paleolithic need say little about the
degree of human biological population continuity. Second, the
geographical location and temporal position of the transition
need not constrain the degree to which there was admixture
between local archaic and early modern human populations.
Third, it is inappropriate to apply a species distinction with
strict implications of reproductive isolation to the Neandertals
versus early modern humans. And fourth, hypotheses (60–63)
that full population replacement of late archaic humans by
early modern humans took place everywhere outside of the
(presumably African) core area of modern humans can be
 Anthropology: Duarte et al. Proc. Natl. Acad. Sci. USA 96 (1999) 7609

rejected. It is therefore necessary to go beyond the categorical 18. Tompkins, R. L. & Trinkaus, E. (1987) Am. J. Phys. Anthropol.
models that have been used commonly to characterize the late 73, 233–239.
archaic to early modern human and the Middle to Upper 19. Mallegni, F. & Trinkaus, E. (1997) J. Hum. Evol. 33, 651–668.
20. Tillier, A. M. (1987) in Préhistoire de Poitou-Charentes, ed.
Paleolithic transitions and to recognize the complex regional,
Vandermeersch, B. (Comité des Travaux Historiques et Scien-
temporal, human biological, and cultural processes as well as tifiques, Paris), pp. 201–206.
historical trajectories that took place. 21. Matiegka, J. (1934) Homo Předmostensis, Fosilnı́ Človĕk z Před-
mostı́ na Moravĕ I (České Akad. Vĕd a Umĕnı́, Prague).
22. McCown, T. D. & Keith, A. (1939) The Stone Age of Mount
CONCLUSION Carmel II (Clarendon, Oxford).
The early Upper Paleolithic human burial at Lagar Velho, 23. Tillier, A. M. (1981) C. R. Acad. Sci. Ser. II 293, 725–727.
Portugal, provides evidence of the burial practices and human 24. Schutkowski, H. (1993) Am. J. Phys. Anthropol. 90, 199–205.
25. Stefan, V. H. & Trinkaus, E. (1999) Bull. Mém. Soc. Anthropol.
biology of these Late Pleistocene human populations. The Paris, 10, 87–118.
burial, with pierced shell ornamentation and red ochre, is 26. Stefan, V. H. & Trinkaus, E. (1998) J. Hum. Evol. 34, 443–468.
similar to Gravettian ones elsewhere in Europe. The facial and 27. Houston, C. S. & Zaleski, W. A. (1967) Radiology 89, 59–66.
appendicular remains provide a morphological mosaic, with 28. Trinkaus, E. & Ruff, C. B. (1996) J. Hum. Evol. 30, 299–314.
the mandibular mentum osseum and dental proportions, as 29. Tardieu, C. & Trinkaus, E. (1994) Am. J. Phys. Anthropol. 95,
well as mandibular ramal, antebrachial features, and pubic 183–195.
size, aligning the specimen with European early modern 30. Anderson, J. Y. & Trinkaus, E. (1998) J. Anat. 192, 279–285.
31. Trinkaus, E. (1984) Curr. Anthropol. 25, 509–514.
humans. The mastoid region is intermediate; however, the 32. Holliday, T. W. (1997) J. Hum. Evol. 32, 423–447.
femorotibial length proportions and lower limb hypertrophy 33. Schultz, A. H. (1923) Am. J. Phys. Anthropol. 4, 389–399.
reflect arctic body proportions, and these features, along with 34. y’Edynak, G. (1978) Am. J. Phys. Anthropol. 45, 569–574.
the mandibular symphyseal retreat and thoracohumeral mus- 35. Vlček, E. (1973) J. Hum. Evol. 2, 537–544.
cle hypertrophy, indicate affinities to the Neandertals. This 36. Madre-Dupouy, M. (1992) L’Enfant du Roc de Marsal (Centre
mosaic indicates admixture between late archaic and early National de la Recherche Scientifique, Paris).
modern humans in Iberia, refuting hypotheses of complete 37. Golovanova, L. V., Hoffecker, J. F., Kharitonov, V. M. &
replacement of the Neandertals by early modern humans and Romanova, G. P. (1999) Curr. Anthropol. 40, 77–86.
38. Ruff, C. B. (1994) Yearb. Phys. Anthropol. 37, 65–107.
underlining the complexities of the cultural and biological 39. Trinkaus, E. (1983) The Shanidar Neandertals (Academic, New
processes and events that were involved in modern human York).
emergence. 40. Trinkaus, E. (1997) Proc. Natl. Acad. Sci. USA 94, 13367–13373.
41. Trinkaus, E., Ruff, C. B., Churchill, S. E. & Vandermeersch, B.
We are grateful to the many curators who have permitted analysis (1998) Proc. Natl. Acad. Sci. USA 95, 5836–5840.
of human remains in their care and to the L. S. B. Leakey, National 42. Trinkaus, E., Stringer, C. B., Ruff, C. B., Hennessy, R. J.,
Science, and Wenner–Gren Foundations for related support. The Roberts, M. B. & Parfitt, S. A. (1999) J. Hum. Evol., in press.
Lagar Velho project is carried out under the auspices of the Instituto 43. Trinkaus, E. & Ruff, C. B. (1999) J. Archaeol. Sci., in press.
Português de Arqueologia, Ministry of Culture, Portugal. 44. Ruff, C. B., Walker, A. & Trinkaus, E. (1994) Am. J. Phys.
Anthropol. 93, 35–54.
1. Vega Toscano, L. G. (1990) Mem. Mus. Prehist. Ile France 3, 45. Trinkaus, E. & Rhoads, M. L. (1999) J. Hum. Evol. in press.
169–176. 46. Trinkaus, E. (1999) in Gibraltar and the Neanderthals, eds.
2. Villaverde, V. & Fumanal, M. P. (1990) Mem. Mus. Prehist. Ile Stringer, C. B. & Findlayson, J. C. (Oxbow, Oxford), in press.
France 3, 177–183. 47. Matiegka, J. (1938) Homo Předmostensis, Fosilnı́ Človĕk z Před-
3. Zilhão, J. (1993) in El Origen del Hombre Moderno en el Suroeste mostı́ na Moravĕ II (České Akad. Vĕd a Umĕnı́, Prague).
de Europa, ed. Cabrera, V. (Univ. Nacl. Educ. Distancia, Ma- 48. Churchill, S. E. (1994) Ph. D. Thesis (Univ. New Mexico,
drid), pp. 127–145. Albuquerque, NM).
49. Heim, J. L. (1982) Les Enfants Néandertaliens de La Ferrassie
4. Hublin, J. J., Barroso Ruiz, C., Lara, P. M., Fontugne, M. &
(Masson, Paris).
Reyss, J. L. (1995) C. R. Acad. Sci. Ser. II 321, 931–937.
50. Holliday, T. W. (1997) Am. J. Phys. Anthropol. 104, 245–258.
5. Garcia Sanchez, M. (1986) Homenaje a Luis Siret, (Consejeria
51. Mengel, R. M. (1971) J. Mammal. 52, 316–336.
Cult. Junta Andalucia, Granada, Spain), pp. 49–56.
52. Nagel, U. (1973) Folia Primatol. 19, 104–165.
6. Zilhão, J. (1997) O Paleolı́tico Superior da Estremadura Portu-
53. Markarjan, D. S., Isakov, E. P. & Kondalov, G. I. (1974) J. Hum.
guesa (Colibri, Lisbon). Evol. 3, 247–255.
7. d’Errico, F., Zilhão, J., Julien, M., Baffier, D. & Pelegrin, J. 54. Cheverud, J. M., Jacobs, S. C. & Moore, A. J. (1993) Am. J.
(1998) Curr. Anthropol. 39, S1–S44. Primatol. 31, 23–39.
8. Zilhão, J. (1998) in XIII International Congress of Prehistoric and 55. Peres, C. A., Patton, J. L. & da Silva, N. F. (1996) Folia Primatol.
Protohistoric Sciences: Proceedings, eds. Facchini, F., Palma di 67, 113–124.
Cesnola, A., Piperno, M. & Peretto, C. (Abaco, Forli, Italy), Vol. 56. Bräuer, G. (1984) in The Origins of Modern Humans, eds. Smith,
2, 299–312. F. H. & Spencer, F. (Liss, New York), pp. 327–410.
9. Zilhão, J. (1999) in Gibraltar and the Neanderthals, eds. Stringer, 57. Smith, F. H. & Trinkaus, E. (1991) in Aux Origines d’Homo
C. B. & Findlayson, J. C. (Oxbow Books, Oxford), in press. sapiens, eds. Hublin, J. J. & Tillier, A. M. (Presses Univ. France,
10. Svoboda, J., Ložek, V. & Vlček, E. (1996) Hunters Between East Paris), pp. 251–290.
and West (Plenum, New York). 58. Hublin, J. J., Spoor, F., Braun, M., Zonneveld, F. & Condemi, S.
11. Aldhouse-Green, S. & Pettitt, P. B. (1998) Antiquity 72, 756–772. (1996) Nature (London) 381, 224–226.
12. Giacobini, G. (1999) Congr. Prehist. France 24, 29–39. 59. Krings, M., Stone, A., Schmitz, R. W., Krainitzid, H., Stoneking,
13. Zilhão, J. (1991) Etud. Rech. Archeol. Univ. Liege 42, 485–501. M. & Pääbo, S. (1997) Cell 90, 1–20.
14. Moorrees, C. F. A., Fanning, E. A. & Hunt, E. E. (1963) J. Dent. 60. Stringer, C. B. & Andrews, P. (1988) Science 239, 1263–1268.
Res. 42, 1490–1502. 61. Vigilant, L., Stoneking, M., Harpending, H., Hawkes, K. &
15. Smith, B. H. (1991) in Advances in Dental Anthropology, eds. Wilson, A. C. (1991) Science 253, 1503–1507.
Kelley, M. A. & Larsen, C. S. (Wiley, New York), pp. 143–168. 62. Harpending, H. C., Batzer, M. A., Gurven, M., Jorde, L. B.,
16. Anderson, M., Messner, M. B. & Green, W. T. (1964) J. Bone Rogers, A. R. & Sherry, S. T. (1998) Proc. Natl. Acad. Sci. USA
Joint Surg. Am. 46, 1197–1202. 95, 1961–1967.
17. Brabant, H. & Twiesselmann, F. (1967) Bull. Group. Int. Rech. 63. Cavalli-Sforza, L. L. (1998) Proc. Natl. Acad. Sci. USA 95,
Sci. Stomatol. 10, 5–180. 11501–11503.